Similar but different: Interviewing monozygotic twins discordant for musical practice

Participants

The participants were recruited from a sample of 11,543 twin individuals born between 1959 and 1985 who participated in an extensive web survey during the fall of 2012 (Mosing et al., 2014). The participating twins were all part of the STAGE cohort of the Swedish Twin Registry (Lichtenstein et al., 2006; Magnusson et al., 2013). In the Swedish Twin Registry, zygosity is determined based on questions about within-pair similarities (Lichtenstein et al., 2002; Lichtenstein et al., 2006). In 27% of the twins zygosity has also been determined with genotyping, showing that the questionnaire-based zygosity determination was correct for more than 98% of twins (Lichtenstein et al., 2002; Lichtenstein et al., 2006). In the present study, questionnaire data from this web survey on music practice, personality (openness to experience), and musical flow proneness were used (see below).

In total, 1,312 complete monozygotic twin pairs participated in the web survey. From these a smaller subsample of twins were invited to participate in the present study, which involved the interviews presented here as well as several other measurements (see Procedure section) that typically took place during the course of two consecutive days, e.g. a weekend. The specific inclusion criteria were that (1) one of the members of the twin pair played a keyboard instrument (piano, organ, or keyboard) at the time of the web survey while the other one did not; (2) the within-pair difference in total hours of music practice was at least 1,000 hours; (3) both twins fulfilled general inclusion criteria for magnetic resonance imaging. An initial screening of the web survey practicing data identified 83 monozygotic pairs who potentially could fulfill the inclusion criteria. These were contacted by telephone to gather further information. During the telephone conversation it was found that the majority of the contacted pairs could not be involved in the study, either since at least one of the twins could not for medical reasons participate in magnetic resonance imaging (the results of which are not reported here), or because of unwillingness or lack of time to travel to Stockholm and participate. The final sample consisted of 10 pairs, aged between 31 and 47 years. Six of the pairs were female; four were male. All participants except two had a university level education. None of the participants was a professional musician. The participants were given 2,000 SKR as a monetary compensation for their participation. All participants gave informed consent to participate and the study was approved by the Regional Ethical Review Board in Stockholm (Dnr 2011/570-31/5; 2012/788-31/2).

Procedure

The interviews took place at Karolinska Institutet (Stockholm) during the period March 2013 to May 2014. In addition to the semi-structured interview which is the focus of the present report, the twins also participated in structural and functional magnetic resonance imaging, physiological measurements, delivered saliva samples, and performed additional behavioral tests related to music reading and finger force control. Analyses of these measurements will be presented elsewhere. The interviews were performed with each twin separately and the order of the interview (i.e. which twin was interviewed first) was randomized within pairs.

Interview format and method of analysis

The interviews were performed by one of the authors (HE) and were structured into five sections. The questions asked by the interviewer in each section are shown in Table 1. The first section focused on the twin’s own thoughts about why one of them had been more musically engaged than the other. In the second section, this was further explored with a number of specific questions concerning within-pair differences in the musical childhood environment. The third section focused on memories of strong musical experiences that may have influenced the participant’s interest and engagement in music. In the fourth section, the twin was asked to reflect on the perceived significance of music in life and on the possible importance of music for personal well-being and health. The final section concerned the participant’s interest and skill in languages. This section was only included for eight of the 10 pairs. At the end of the interview there was a brief summary discussion. The interviews lasted for 30 to 40 minutes and were tape recorded. Following Ahrne and Svensson (2011) the interviews were transcribed verbatim during the analysis phase. Keywords and key sentences were marked in the text and on the basis of these, response categories were developed for each question (see Results section).

Web survey data

Web survey data

The age of the participants ranged between 31 and 47 years (M = 37.4, SD = 6.0). As predicted, openness to experience was higher [Cohen’s d = 0.90 (95% CI: 0.18–1.61), paired t-test, t(9) = 2.84, one-tailed p = .01] in the playing twins (M = 3.9, SD = .73) than in the non-playing twins (M = 3.3, SD = .67). Likewise, the difference in musical flow proneness between the playing twins (M = 3.63, SD = .40) and their non-playing co-twins (M = 3.16, SD = .45) was statistically significant [Cohen’s d = 0.95 (95% CI: 0.24–1.67), paired t-test, t(9) = 3.01, one-tailed p = .007]. Since parametric tests could potentially be problematic with smaller samples, we also verified that the results remained significant using the non-parametric Wilcoxon Matched Pairs Test. This was the case both for openness to experience (Z = 1.99, one-tailed p = .02) and music flow proneness (Z = 2.31, one-tailed p = .01).

Interview data

Section 1 – perceived reasons for the discordance in musical engagement

In section 1, a general question about possible reasons for their musical discordance (Table 2) was asked, and the twins were allowed to reflect freely on this. Responses in section 1 are summarized in Table 2. Altogether, 17 categories were identified among the responses. By far the most common category, observed in eight of the 10 pairs, was that the playing twin found music more interesting and enjoyable than his or her co-twin. We illustrate this response category with quotes from three playing and two non-playing twins:

I’ve had the interest to continue with it and keep it alive. It’s more of a question why he hasn’t had the same wish… It surprises me a bit that NN [co-twin] hasn’t continued. Yes, that’s how I think. (playing twin)

We had a piano at home, and I thought it was fun to sit there and sing along. But there we were very different. So when it was Christmas I got to play and show off, but she was not at all interested in that. (playing twin)

Well, I thought it was more fun. So that I wanted to continue playing. It was not so that our parents forced or controlled us… no, I thought it was more fun. (playing twin)

I started to play a bit, but then I grew tired of it and did not think it was as enjoyable as my brother thought. Basically, I stopped practicing. (non-playing twin)

The passion to play hasn’t been there, I haven’t felt any drive. That came gradually. It gave her more than me. When I began puberty, this thing with the piano was not the same any more. (non-playing twin)

OPEN IN VIEWER

Table 2.

Categorized responses in section 1 (perceived reasons for the discordance in music practice) in each twin pair.

	Twin pair																				N(pairs)	N(P)	N(N)
	1		2		3		4		5		6		7		8		9		10
Response category	P	N	P	N	P	N	P	N	P	N	P	N	P	N	P	N	P	N	P	N
P found music more interesting and enjoyable and showed more zeal than N	x				x	x		x	x	x	x	x		x			x	x	x	x	8	6	6
Non-musical interests were more important for N										x	x	x						x			3	1	3
P had more talent than N from start		x									x	x									2	1	2
N started a family while P was/is still single					x	x						x									2	1	2
Only P had access to a piano as an adult			x	x			x														2	2	1
P and N differed in social activities and interests during childhood and adolescence	x														x						2	2	0
P practiced more than N from start															x			x			2	1	1
P and N competed about time at the piano																		x			1	0	1
N did not enjoy the musical genre that was performed																x					1	0	1
N chose a creative profession and felt less need for piano playing than P				x																	1	0	1
The piano was placed in the room of P which made it less accessible for N			x																		1	1	0
P had more musically interested acquaintances than N								x													1	0	1
P liked their common piano teacher more than N did	x																				1	1	0
For P more than N, making music was a way to express religious faith													x								1	1	0
P had generally more vocational interests than N													x								1	1	0
N played only for social reasons																				x	1	0	1
I do not know																			x	x	1	1	1

Note. P = playing twin, N = non-playing twin. The columns N(pairs), N(P), and N(N) summarize how many pairs, playing twins, and non-playing twins, respectively, produced a response in each category. Responses are sorted in descending order of frequency of the response category.

Within three of these eight pairs, one or both twins also explicitly mentioned that the non-playing twin had a higher level of interest in non-musical activities (Table 2). A perceived difference in musical talent was, in contrast, only reported in two pairs. The remaining response categories consist of a wide range of possible reasons for the discordance suggested by the respondents in different pairs. Each of these response categories was seen only in one or two pairs (Table 2). The response categories are not mutually exclusive and all pairs provided more than one explanation for their discordance. In one pair, both twins stated explicitly that they were not sure why they developed different musical interests.

Section 2 – specific within-pair differences in musical childhood environment

In section 2, the discussion in section 1 was followed up with a number of specific questions concerning differences between the twins in childhood factors that could be potentially relevant for musical engagement.

The first question concerned differences in music listening habits (Table 3). Responses were grouped into four categories. As seen in Table 3, in six pairs the twins reported that the listening habits were identical or similar. In three pairs both twins reported that the playing twin listened more frequently to music, and in one pair both twins stated that they did not know whether there was a difference. The two twins in one pair which differed in reported music listening habits expressed it as follows:

From age 11–12 on I would say there was quite a big difference. That I, as I perceived it, always listened more to music. And searched… been more curious about finding new music. Bought more too. (playing twin)

It was my sister [who listened more]. And then it was so that first we shared a room, later she had the room next to me, and what happened was also that if she already played music, I wouldn’t even think of putting on a record or so because she had already put on something. It has been like that a lot in my life, because later on I met a man who was interested in music, and then it ended up in the same way, that he puts on something, before I even think about whether I want to or not. (non-playing twin)

OPEN IN VIEWER

Table 3.

Categorized responses in section 2, differences in music listening habits.

	Twin pair																				N(pairs)
	1		2		3		4		5		6		7		8		9		10
Response category	P	N	P	N	P	N	P	N	P	N	P	N	P	N	P	N	P	N	P	N
No difference between P and N				x	x	x					x	x	x	x			x	x			5
P and N had similar habits	x		x																		2
P listened more than N		x					x	x	x	x					x	x					4
Do not know																			x	x	1

Note. P = playing twin, N = non-playing twin. The column N(pairs) indicates in how many pairs a given response category was observed.

The second question concerned whether the twins had the same or different music teachers. In nine of 10 pairs both twins responded that they had the same music teacher during childhood. In the remaining pair, both twins reported that they played different instruments and thus had different teachers.

Third, the participant was asked whether he or she remembered any within-pair difference in terms of the musical engagement of the childhood peers of the twins. In nine of 10 pairs both twins reported that they did not recall any difference in the musical interests of their respective peers. In one pair, the playing twin reported that she had one more musically interested peer than her co-twin, and that this may have stimulated her musical interest:

… the guy I was around most, he played the piano. Of course that may have had an influence. When we were at his place, both his brother and both parents played. There was a lot of music there, and that may have influenced me. (playing twin)

However, her co-twin did not mention this, stating rather that the two sisters mixed with the same group of peers.

The fourth question concerned differences in participation in musical ensemble playing. None of the pairs reported any within-pair difference in this regard. In five pairs, both twins participated in musical ensembles; in the other five pairs, none of the twins did.

Fifth, the participant was asked if he or she experienced any difference in the engagement of their parents in the musical training of the two twins. In all pairs, both twins explicitly stated that there was no difference in parental support. However, in one pair, the non-playing twin mentioned that the father practiced with both twins but typically started with his brother (the playing twin):

I don’t know how much truth there is in it, but in the beginning Dad was the one who was with and practiced with us, and you know it’s difficult to practice with both at the same time, so you start with one [laughter] and I think that it always was… NN [co-twin] is a few minutes older you know, twin 1, and I am 2, and I think that he most often began with NN. And when you do the same thing again, then it might not be as fun. You see how I mean? Because he had to do everything twice, Dad [laughter]. (non-playing twin)

In another pair, the non-playing twin suggested that the mother, who was musically interested, was an important role-model for her sister, whereas the less musically interested father was a primary role-model for herself:

Generally I imitated my Dad, and my sister imitated our Mom. And she was more driving with this thing to play the piano. (non-playing twin)

Finally, the participants were asked about possible differences between the twins in terms of public music performances. In one pair, both twins indicated that the playing twin had made somewhat more frequent public appearances during childhood. However, in the remaining pairs the twins did not report any differences in public performances. In all cases the twins stated that they had performed publicly only infrequently, or not at all and that, if at all, they had only performed in smaller venues, for example privately or in the music school.

To summarize the findings from section 2, in three pairs the playing twin was reported to have listened more to music. Apart from this, the specific questions in this section did not reveal any striking differences between the twins; that is, in almost all cases they had the same music teacher, and reported no differences in the musical interests of their peers, the engagement of the parents in their musical education, or participation in ensemble playing and public performances.

Section 4 – the meaning of music in life and for well-being

In section 4, the twins were asked to reflect on the significance of music for them, both in life in general and more specifically for their personal well-being and health. Responses in this section were classified into 10 categories, which are summarized in Table 4, sorted in descending order of frequency. As can be seen, the three most common response categories were relaxation and meditation (five pairs), enjoyment (four pairs), and general statements expressing that music is an important part of life (four pairs). The relaxation/meditation category can be illustrated by the following quotes from twins in two different pairs:

Above all it means relaxation at home. It is about recovery, and to get your mind off something, some job. That’s the main thing. (playing twin)

For me it’s a source of relaxation and a way to clear away other thoughts. Music is something you experience in the present moment. (non-playing twin)

OPEN IN VIEWER

Table 4.

Categorized responses in section 4 (the perceived meaning of music in life and for well-being and health) in each twin pair.

	Twin pair																				N(pairs)	N(P)	N(N)
	1		2		3		4		5		6		7		8		9		10
Response category	P	N	P	N	P	N	P	N	P	N	P	N	P	N	P	N	P	N	P	N
Relaxation and meditation			x		x		x									x	x				5	4	1
Enjoyment							x					x	x					x			4	2	2
Being an important part of life								x				x	x						x		4	2	2
Forming bonds to other people			x						x							x					3	2	1
Finding consolation and refuge				x									x		x						3	2	1
Finding your identity			x						x												2	2	0
Express your creativity					x								x								2	2	0
Making music with your children	x																		x		2	2	0
Controlling your emotional state		X													x						2	1	1
Enhancing positive self-image	x																				1	1	0
Promoting flow experiences											x										1	1	0
Promoting general knowledge										x											1	0	1
Promoting happiness in social contexts						x															1	0	1
No particular importance																				x	1	0	1
Number of responses	2	1	3	1	2	1	2	1	2	1	1	2	4	0	2	2	1	1	2	1	–	21	11

Note. P = playing twin, N = non-playing twin. The columns N(pairs), N(P), and N(N) summarize how many pairs, playing twins, and non-playing twins, respectively, produced a response in each category.

The remaining categories cover a wide range of themes, that involve both social interactions with others (forming bonds to other people, making music with your children, promoting happiness in social contexts) and the expression of yourself as an individual (finding your identity, express your creativity, enhancing positive self-image, promoting general knowledge).

Overall the responses of the playing twins in this section were richer and more varied than those of the non-playing co-twins. The last row of Table 4 shows the total number of responses in each column. Notably, the playing twins provided a total of 21 responses in all categories, as compared to 11 for the non-playing twins. At the pairwise level, the playing twin used more categories than the non-playing co-twin in seven pairs. In two pairs, the two twins used the same number of categories, and in one pair the non-playing used two categories while the playing twin used one. We end this section with a striking quote from a playing twin in the response category “promoting flow experiences”:

Oh, [music] means enormously much! [laughter] I am the happiest at concerts or when I can sit down and play, and things just flow, so to speak. That is what you really long for. When you really get into that flow, that is. (playing twin)

Psychological differences in twins discordant for music

As summarized in the introduction, active musical engagement is related to a multitude of factors that involve genetically influenced traits of the individual as well as a host of environmental variables, including parental socioeconomic status and education. A main aim of the present study was to identify factors that may influence musical engagement over and above both genetic constitution and common family environment, by studying monozygotic twins that were discordant for piano playing.

One key finding of the study is that the playing twins tended to have a stronger psychological relation to music than their co-twins. This was revealed in several ways. Results from the web survey data showed that the twins differed significantly in musically relevant traits relating both to personality, i.e. openness to experience, and intrinsic motivation, operationalized as the proneness to have psychological flow experiences during music. The effect sizes of these differences were large (Cohen’s d = 0.90–0.95). These findings were expected, given the earlier studies that have shown phenotypic associations between music engagement and openness to experience (Butković et al., 2015; Corrigall & Schellenberg, 2015; Corrigall et al., 2013; Dyce & O’Connor, 1994; Gillespie & Myors, 2000; Greenberg et al., 2015; Kaufman et al., 2016; Müllensiefen et al., 2014), as well as between flow proneness and engagement in creative domains (Csikszentmihalyi, 1997). In particular, recent twin modeling studies on the whole cohort from which the current sample was recruited have shown that the associations between both these traits and music practice are driven by non-genetic as well as genetic influences (Butković et al., 2015).

Observations from the semi-structured interviews provide a more fine-grained picture of what these differences in interests and motivation may imply in the daily life of the individual. In section 1, by far the most common response was that the playing twin from early on showed a higher interest for music, enjoyed music more and showed a higher level of motivation, i.e. experienced more zeal when practicing the piano. In several of the pairs, the playing twin was also said to listen more to music already from an early age (section 2). Interestingly, the responses when the twins were asked to reflect on the meaning of music for them for life in general and well-being (section 4) were considerably richer and more varied from the playing than from the non-playing twins. While music as a means for relaxation and meditation was the most common category, it was also clear that music could be important for social life as well as personal development. Among social functions of music, forming bonds to other people and making music with one’s children were mentioned. Interestingly, group performance is commonly found in musical cultures all over the world (Savage, Brown, Sakai, & Currie, 2015), and according to some evolutionary theories of the origins of music, increasing social cohesion could be a primary function of musical activities (Kirschner & Tomasello, 2010; Savage et al., 2015).

Categories relating more to individual development included functions relating to the regulation of emotion, such as enjoyment, finding consolation and refuge, controlling your emotional state, and enhancing a positive self-image, but also functions relating more to cognitive development such as building a personal identity, finding flow, and the expression of creativity.

Generally, most categories were found in non-playing as well as playing twins, although the latter group used more categories overall. Interestingly, however, the last three categories above – i.e. building a personal identity, finding flow, and the expression of creativity – were represented only in the answers of playing twins, in five different pairs (see pairs 2, 3, 5, 6, and 7; Table 2). This suggests that for these playing twins, more than for their non-playing co-twins, music was identified as an activity of fundamental importance for self-realization and personal identity development (Erikson, 1980; Maslow, 1970). For a recent discussion – from the perspective of eudaimonistic identity theory – of how participation in activities involving personal expressiveness and flow can promote personal identity, see Mao, Roberts, and Bonaiuto (in press). A couple of more anecdotal observations appear to be in line with the conclusion that music played a role for identity formation and cognitive growth in the playing twins. When reflecting on musical memories (section 3), one playing twin mentioned the desire to role model the local organist as an important driving force for her own practicing. Another playing twin mentioned strong memories relating to the discovery of new musical styles and genres.

Why did the twins become discordant for musical engagement?

An intriguing question is why the twins became so discordant in terms of musical interests and engagement. The twins in all pairs had grown up together, meaning that they shared not only all genetic factors but also a common family environment. Parental socioeconomic status, a consistent predictor of cultural engagement (Corrigall & Schellenberg, 2015; Corrigall et al., 2013; Schellenberg, 2006, 2011; Sergeant & Thatcher, 1974; Theorell et al., 2015) was obviously the same within pairs. Furthermore, the interviews (sections 2 and 5) showed that the twins, with rare exceptions, experienced no within-pair differences in a number of other variables that potentially could influence musical engagement; that is, the music teacher, the musical engagement of peers, ensemble playing, parental support, public performances, and their interest and aptitude for languages. In relation to the latter point, one can also note that the associations between music practice, on the one hand, and musical aptitude (Mosing et al., 2014) and general intelligence (Mosing et al., 2016), on the other hand, have been analyzed using twin modeling in the whole cohort from which the current sample was selected. These analyses demonstrated that the phenotypic associations were essentially driven by genetic factors, with the consequence that no systematic relations between music practice and general or music specific cognitive abilities were seen within monozygotic twin pairs.

Taken together, the data provide no indication that the observed discordance was caused by any systematic environmental factor that was consistent across twin pairs. On the contrary, not less than 15 of the response categories in section 1, where the twins were asked to reflect on possible reasons for the discordance, were recorded in only one or two pairs. An examination of these categories reveals a strikingly wide range of possible reasons for the discordance suggested by the respondents. These include differences in access to a piano, either while growing up or in adult life; differences in attitudes towards the musical genres performed; differences in the perceived need for a creative hobby; differences in social activities and musical interests among friends; different feelings about the common music teacher; and different attitudes towards music as an expression of religious faith. To this list can be added anecdotal observations from sections 2 and 3: one playing twin mentioned that the musical engagement of one of her peers may have been an influence on her own playing, although no relevant differences in peer interests were recalled by the other nine pairs. In another pair, the musically interested mother may have been a more important role model for the playing twin than for the non-playing twin. In a third pair, the playing twin was reported to play more in public during childhood. In section 3, strong musical memories included unique events such as the negative memory of a failed public performance in one non-playing twin.

Of course this kind of self-report data has to be interpreted with great caution. We have no information on the reliability and validity of the provided responses. Furthermore, even if the twins provide an accurate recollection of past events the issue of causality remains unclear: even in those cases where a remembered event in retrospect appears to have had a causal impact, this subjective experience is obviously not evidence that this is in fact the case. Nevertheless, the overall impression from the data is that the environmental factors which potentially may have had a causal influence on the twins’ musical discordance are multifarious and unsystematic in nature. Notably, the interviews gave no hint of environmental influences on the discordance that were consistent across pairs.

In behavior genetic terms, the environmental factors underlying the discordance will by definition all belong to non-shared environment, an umbrella term for the set of non-genetic influences that tend to make siblings in the same family different from one another (Plomin & Daniels, 1987; Plomin, DeFries, Knopik, & Neiderhiser, 2013). One should note that the definition of non-shared environment is based on its effect: an environmental variable that is objectively shared by two twins in a family, in the sense that they receive equal exposure to it, would still belong to non-shared environment if its effect is to make the twins more different from one another (Turkheimer & Waldron, 2000). To take an example relevant to the present data, the musical activities of one parent could be a non-shared environmental influence if this parent serves as a role-model for one twin, but not the other. Turkheimer and Waldron (2000) performed a meta-analysis of studies reporting the effects of specific, measured non-shared environmental variables (e.g. differential parental behavior and differences in interactions with peers, teachers, and siblings) on various outcomes (adjustment, personality, and cognitive ability). In general, estimated effects were very small, with a median effect of r² = .010 for studies using genetically informative designs. One much-discussed possibility is thus that non-shared environmental variance generally is the cumulative result of a large number of events, each of which only has a small and unsystematic effect on the phenotype (Plomin & Daniels, 1987; Turkheimer & Waldron, 2000). An obvious limitation of the present study is the small sample size, but the failure to identify any consistent external factors that could explain the observed discordance in musical behaviors in the quite extensive interview material seems in line with this account of non-shared environment.